The thiol-disulfide exchange activity of AtPDI1 is involved in the response to abiotic stresses

Arabidopsis protein disulfide isomerase 1 (AtPDI1) has been demonstrated to have disulfide isomerase activity and to be involved in the stress response. However, whether the anti-stress function is directly related to the activities of thiol-disulfide exchange remains to be elucidated. In the present study, encoding sequences of AtPDI1 of wild-type (WT) and double-cysteine-mutants were transformed into an AtPDI1 knockdown Arabidopsis line (pdi), and homozygous transgenic plants named pdi-AtPDI1, pdi-AtPDI1m1 and pdi-AtPDI1m2 were obtained. Compared with the WT and pdi-AtPDI1, the respective germination ratios of pdi-AtPDI1m1 and pdi-AtPDI1m2 were significantly lower under abiotic stresses and exogenous ABA treatment, whereas the highest germination rate was obtained with AtPDI1 overexpression in the WT (WT- AtPDI1). The root length among different lines was consistent with the germination rate; a higher germination rate was observed with a longer root length. When seedlings were treated with salt, drought, cold and high temperature stresses, pdi-AtPDI1m1, pdi-AtPDI1m2 and pdi displayed lower survival rates than WT and AtPDI1 overexpression plants. The transcriptional levels of ABA-responsive genes and genes encoding ROS-quenching enzymes were lower in pdi-AtPDI1m1 and pdi-AtPDI1m2 than in pdi-AtPDI1. Taken together, these results clearly suggest that the anti-stress function of AtPDI1 is directly related to the activity of disulfide isomerase.


Background
As a thiol-disulfide oxidoreductase, protein disulfide isomerase (PDI) catalyzes the oxidative folding and isomerization of nascent polypeptide chains [1][2][3]. Some PDIs also display chaperone or anti-chaperone activity [4,5]. Typical PDI, for example, human PDI (hPDI), contains two thioredoxin-like domains, a and a′, linked by domains b and b′ [6]. The enzymatic activities of PDI are directly related to the CXXC motif in the a and a′ domains, respectively, and mutations of the conserved cysteines in the motifs lead to loss of enzymatic activities [7][8][9]. Extensive investigations have demonstrated that animal PDIs, such as hPDI, are located in the endoplasmic reticulum (ER) with a carboxyl-terminal KDEL sequence as the retention signal [6]. Under adverse conditions, the process of correct protein folding is disturbed, which leads to ER stress and the unfolded protein response (UPR), finally resulting in cell death [10]. When the human body is threatened by pathogens, the expression of PDIs is upregulated, keeping unfolded or misfolded proteins correctly folded with functional Open Access *Correspondence: shizhong@sdau.edu.cn; qhsun@sdau.edu.cn † Ying Lu and Li Yuan contributed equally to this work. 1 College of Life Science, State Key Laboratory of Crop Biology, Shandong Agricultural University, Taian, Shandong 271018, People's Republic of China Full list of author information is available at the end of the article conformation [11][12][13][14]. Furthermore, investigations demonstrated that inhibition of hPDI caused the accumulation of unfolded or misfolded proteins [15,16]. Generally, the formation of a native conformation of the peptide chain occurs with the help of molecular chaperones and is then oxidized to form disulfide bonds. When the activity of oxidoreductase was blocked, hPDI could also act as a chaperone, binding to misfolded or unfolded proteins and facilitating proper folding to mitigate the damage under stress conditions, which means that the chaperone activity is independent of its redox active site [17].
Homologous proteins of hPDI have also been found in different plants, such as Oryza sativa, Triticum aestivum L., Brassica rapa ssp. pekinensis, Zea mays and Arabidopsis thaliana [18][19][20][21]. Arabidopsis contains at least 12 PDI members, and they are divided into 3 groups based on polypeptide length, the presence of signal peptides and ER retention signals, and the composition of thioredoxin domains [19,[22][23][24][25]. AtPDI1 (encoded by At3g54960) belongs to group II, having a typical PDI structure with 4 modular domains (a, b, b′, a′) and a CXXC motif in each a and a′ domain, respectively [22]. In contrast to hPDI, an additional signal peptide is found in the amino terminus that targets chloroplasts, although AtPDI1 also contains the ER retention signal (KDEL sequence) at the carboxyl terminus [25]. Experiments showed that AtPDI1 was distributed in both organelles [23,24], suggesting that it may play roles in both the ER and chloroplasts.
As one of the most important organelles, chloroplasts contain a considerable proportion of proteins whose activities can be changed by thiol-disulfide exchange [24,26]. In Arabidopsis chloroplast stroma, m-type thioredoxin (Trx m) is the most abundant type of protein and it mainly catalyzes the reduction of disulfide bonds involved in photosynthetic carbon assimilation and accumulation in photosystem II [26], while the roles of oxidation or the formation of disulfide bonds remain uncertain. AtPDI1 is most likely to oxidize the reduced thiol group in chloroplast stroma, catalyzing the exchange of thiol-disulfide [24]. In the ER, similar to animal PDI, AtPDI1 is also the key enzyme catalyzing oxidative folding of secretory proteins and is also related to ER stress during the UPR [22].
Since plants are immobile, they are easily subjected to abiotic stresses. Similar to biotic stresses in humans (pathogen infection), abiotic stresses also lead to unfolded and misfolded proteins, causing loss of functional activities [27]. Recent studies have shown that some plant PDI members are involved in abiotic stresses [28][29][30][31][32]. Among 12 Arabidopsis PDIs, AtPDI1 is strongly induced by abiotic stresses and exogenous abscisic acid (ABA) [28,33,34]. Our previous study showed that the pdi (AtPDI1 T-DNA insertion knockdown mutant line) was more sensitive to abiotic stresses than WT Arabidopsis. Expression of recombinant AtPDI1 significantly improved the resistance to abiotic stresses in E. coli [28]. To clarify whether the anti-stress ability of AtPDI1 is related to its catalytic activity, here, the conserved cysteines in the CXXC motif(s) were mutated, and the pdi mutant line was used as the transgenic plant host. Two double mutants, AtPDI1 m1 (AtPDI1 C128AC131A ) and AtPDI1 m2 (AtPDI1 C467AC470A ), and wild-type AtPDI1 (AtPDI1), were constructed with the plasmids and successfully transformed into the pdi line. The transgenic lines were named pdi-AtPDI1 m1 , pdi-AtPDI1 m2 and pdi-AtPDI1. Then, the above lines, together with pdi, the WT and an overexpression line of the WT (WT-AtPDI1), were all treated with different stresses to study whether the thiol-disulfide exchange activity of AtPDI1 is involved in the response to abiotic stresses.

Results
In contrast to wild-type AtPDI1, mutations in AtPDI1 could not compensate for the phenotype of pdi under abiotic stresses during germination The seeds of different lines, including those of pdi, WT, WT-AtPDI1, pdi-AtPDI1, pdi-AtPDI1 m1 and pdi-AtP-DI1 m2, were all sown on the same stress medium to check whether the conserved cysteine residues have critical roles in the response to abiotic stresses. Two statistical methods, radicle appearance and cotyledon extension, were used to investigate the process of germination. Under normal conditions (1/2 MS medium), no difference in germination rate or germination speed was observed among the 6 lines when the radicle appearance was taken as the criterion, with germination reaching 100% at approximately 3 d (Fig. 1A, B). However, when the seeds were sown on medium containing 150 mM NaCl, 200 mM mannitol or 2 mM H 2 O 2 , significant differences in germination were observed (Fig. 1A, B). Under 200 mM mannitol treatment, the pdi line germinated much slower than the WT line, with an average of 15% vs. 65% germination in 3 d, respectively, whereas the WT AtPDI1 successfully competed for the deficient phenotype of the pdi line; that is, the germination rate of pdi-AtPDI1 was similar to that of the WT, reaching approximately 65% germination in 3 d. Most notably, the germination of pdi-AtPDI1 m1 and pdi-AtPDI1 m2 was quite similar to that of pdi, suggesting that the mutations of the conserved cysteines in the CXXC motif caused AtPDI1 to completely lose its antistress ability during germination. Overexpression of AtPDI1 enhanced the anti-stress ability and increased the germination of WT-AtPDI1 compared with that of the WT (Fig. 1B). In addition, a similar tendency occurred with the 2 mM H 2 O 2 and 150 mM NaCl treatments . When monitoring cotyledon extension during germination under different stresses, a similar tendency among different lines was observed (Fig. 1C).
To assay the difference in root length among the 6 lines, the seeds were sown on 1/2 MS medium until radicle appearance and then transferred to longitudinal stress medium containing NaCl, mannitol or H 2 O 2. As shown in Fig. 2, although little difference was found among the 6 lines under normal conditions, significant differences in root length were observed among them under the abiotic stress treatments. The WT and pdi-AtPDI1 lines had almost the same root length, and overexpression of AtPDI1 (WT-AtPDI1) promoted root length, while the root length of the other transgenic lines (pdi, pdi-AtPDI1 m1 , pdi-AtPDI1 m2 ) was Fig. 2 Root length of WT, pdi, WT-AtPDI1, pdi-AtPDI1, pdi-AtPDI1 m1 and pdi-AtPDI1 m2 lines on medium containing H 2 O 2 , NaCl or mannitol for 10 days, and the control for 7 days. The statistical significance level is shown using different letters. Each experiment was carried out in three biological replicates significantly shorter. The results collectively indicated that the mutations of the cysteines in AtPDI1 inhibited root growth under abiotic stresses (Fig. 2).

Mutations of the active site in AtPDI1 decreased seedling resistance to abiotic stresses
Seedlings of different lines of the same age were treated with salt, osmotic and extreme temperature stresses. When 7-day-old seedlings were watered with 250 mM NaCl for 13 d, little difference among different lines was observed. However, when the treatment lasted for 21 d, significant differences emerged. The WT and pdi-AtPDI1 lines grew well and had a similar etiolation rate (approximately 35%), while the pdi, pdi-AtPDI1 m1 and pdi-AtPDI1 m2 lines turned yellow or even died, with an etiolation rate of approximately 70%, the WT-AtPDI1 plants grew best with an etiolation rate of only 8% (Fig. 3A, Fig. S1A). After PEG treatment for 21 d, the WT-AtPDI1 lines displayed a higher survival rate (80%) than the other lines, whereas the survival rate of the pdi (52%) line was the lowest, followed by the pdi-AtPDI1 m1 (62%) and pdi-AtPDI1 m2 lines (68%) (Fig. S1B).
To test the effects of AtPDI1 on extreme temperature, seedlings were placed in a low-or high-temperature environment. For cold treatment, the seedlings were placed at − 20 °C for 30 min and then placed at room temperature (RT) for 6 h. As shown in Fig. 3C, compared with pdi-AtPDI1 or WT plants, the plants of the pdi, pdi-AtPDI1 m1 and pdi-AtPDI1 m2 lines showed weak tolerance to cold stress, many seedlings exhibited damage and increased leaf juice effusion, with a survival rate of only 63-67% (Fig. S1C), whereas the overexpression line (WT-AtPDI1) could better withstand cold stress, with a survival rate of 90.3% (Fig.  S1C). For the high-temperature treatment, seedlings were treated at 45 °C for 9 h and then placed under normal conditions. After 2 days at RT, WT-AtPDI1 displayed the highest survival rate (100%), the WT fared better than pdi-AtPDI1, with a survival rate of 90%, while most plants of the pdi, pdi-AtPDI1 m1 and pdi-AtPDI1 m2 lines still remained in a withered stated, with a survival rate of approximately 58-75% (Fig. 3D,  Fig. S1D). These results suggested that the conserved cysteines in the active site of AtPDI1 played critical roles in resisting abiotic stresses.
Activities of thiol-disulfide exchange of AtPDI1 affect ABA sensitivity during seed germination ABA signaling is involved in seed germination. To verify whether the change in AtPDI1 activities affects the plant tolerance to ABA, different lines were treated with exogenous ABA to observe the seed germination rate and root length. The results showed that the WT-AtPDI1 line exhibited the highest germination rates and the longest roots among the 6 lines, suggesting that overexpression of AtPDI1 decreased ABA sensitivity. The WT and pdi-AtPDI1 lines had a similar germination rate, which was lower than that of WT-AtPDI1 but higher than that of pdi, pdi-AtPDI1 m1 and pdi-AtPDI1 m2 , and the germination rates of the latter three lines showed no obvious differences. In addition, the difference in root length among these 6 lines coincided with the germination rates (Fig. 4). In addition, the difference in root length among these 6 lines coincided with the germination rates (Fig. 4). The above results indicated that the change in AtPDI1 activities affects the tolerance to ABA.

AtPDI1 functions in the salt stress response may be related to the ABA signaling pathway
To further investigate the relationship between the changes in the stress resistance of the mutant AtPDI1 and the ABA signaling pathway, the expression of key genes related to ABA synthesis (ABA1, ABA2, NCED3) and ABA response (RD29A, KIN1, AnnAt1) was assayed. As shown in Fig. 5, all the above genes were upregulated under NaCl treatment; however, the range of increase was different. The ranges of the expression changes in the abovementioned genes in pdi, pdi-AtP-DI1 m1 and pdi-AtPDI1 m2 were similar and were significantly less than those in the WT and pdi-AtPDI1. For example, the expression of ABA1 was upregulated 12-fold for WT-AtPDI1, approximately 6-fold for WT and pdi-AtPDI1, and less than 3-fold for pdi, pdi-AtP-DI1 m1 and pdi-AtPDI1 m2 . Taken together, the above results preliminarily suggested that AtPDI1 might be involved in the ABA-dependent pathway to enhance resistance to abiotic stresses.

Reactive oxygen species (ROS) include H 2 O 2 , O 2
·-, OH . and singlet oxygen [35,36]. Abiotic stress could result in the production of ROS via the deregulation of electron transport in the chloroplast and mitochondria [35,37]. ROS accumulation in chloroplasts can significantly inhibit the de novo synthesis of D1 protein, thus inactivating the PSII repair process [38]. In addition to causing toxicity, ROS can regulate plant development and act as signaling molecules of abiotic stress [35,37]. Some enzymes, including superoxide dismutase (SOD), catalase (CAT), L-ascorbate peroxidase (APX) and glutathione transferase (GST), have ROS-quenching functions and maintain the balance of ROS in organisms. The transcriptional expression of these enzymes was detected by qRT-PCR. Compared with the lower levels of expression before the abiotic stresses were applied, the expression of these genes was upregulated after the seedlings were treated with 250 mM NaCl. The expression of the genes in the pdi, pdi-AtPDI1 m1 and pdi-AtPDI1 m2 plants was altered less than that in the WT and pdi-AtPDI1 lines, while that in WT-AtPDI1 was upregulated the most (Fig. 6).
In addition, NBT staining was used to measure the accumulation of O 2 ·in different lines under 250 mM NaCl and 300 mM mannitol treatments. More O 2 .accumulated in pdi, pdi-AtPDI1 m1 and pdi-AtPDI1 m2 seedlings with deeper staining than in WT, pdi-AtPDI1 and OE plants (Fig. 7). The accumulation of H 2 O 2 was also checked by DAB staining, but no significant differences among different lines were observed.

Discussion
When the human body is threatened by diseases and other adverse environments, folding-incompetent proteins accumulate, and then ER stress and an associated stress response, called the unfolded protein response (UPR), appear [39]. As important members of the protein folding machinery in the ER, extensive studies have shown that PDIs can catalyze the formation and rearrangement of disulfide bonds of proteins and play a critical role in nascent peptide folding [39][40][41]. hPDI promotes protein folding by catalyzing the exchange of thioldifulfide, which is dependent on the CXXC motif in the a and a′ domains [42].
In plants, PDI is also involved in defense against abiotic stresses, which often disturb cell physiological processes and lead functional proteins to unfold or misfold [20,29]. Similar to hPDI, AtPDI1 has a typical modeling structure of PDI, and some Arabidopsis PDIs are induced to improve protein folding and transport, which subsequently regulate the protein networks under stresses [33]. In this study, we found that the anti-stress activity of AtPDI1 was related to its catalytic activity domain. The pdi line was the most sensitive among the 6 lines under different abiotic stress treatments; however, when AtPDI1 was supplemented in the pdi line (pdi-AtPDI1), the deficient phenotype was restored, and the growth state was consistent with that of the WT. Although pdi-AtPDI1 m1 and pdi-AtPDI1 m2 were successfully expressed in the pdi line (Fig. S1), this could not compensate for the deficiency of pdi, and these lines were still sensitive to stresses, similar to the pdi line. The results suggested that the antistress roles of AtPDI1 are directly related to the activity of thiol-disulfide exchange.
The ABA signaling pathway is an important regulatory mechanism in response to abiotic stress in plants, including salinity, drought and cold [28]. A previous study revealed that the transcript level of AtPDI1 was induced by abiotic stresses and exogenous ABA [28], which suggested that the anti-stress function of AtPDI1 may be related to ABA signaling pathway. In this study, the results showed that the activity of thiol-disulfide exchange affected the resistance to abiotic stress. Moreover, the transcription levels of ABA-responsive genes were substantially higher in WT-AtPDI1 than in WT and the AtPDI1 mutant lines under salt stress conditions, suggesting that AtPDI1 activity is involved in abiotic stress resistance by regulating the expression of ABAresponsive genes.
The change in the activities of AtPDI1 also affected seed sensitivity to ABA, and mutation of AtPDI1 increased sensitivity to ABA, with lower germination and shorter root length under exogenous ABA treatment. Many investigations reported that some genes overexpression exhibited ABA hypersensitive phenotypes during seed germination and improved abiotic stress tolerance through ABA-dependent pathway [42]. However, overexpression of some genes displayed different effects, with ABA -insensitive phenotypes during seed germination and increased stress tolerance through ABA-dependent pathway. These genes come from different plants, such as GhMYB73 [43], TaMYB73 [44], OsMYB3R-2 [45], and AtHD2C [46]. Our results showed that Arabidopsis overexpressing AtPDI1 had a similar phenotype to Arabidopsis overexpressing the above genes, affecting abiotic stress resistance through ABA signaling pathway. However, how AtPDI1 is involved in the ABA pathway to increase abiotic tolerance needs further study.
Abiotic stress can result in the accumulation of excessive ROS, which have been proven to have a negative effect on abiotic stress resistance in plants. Therefore, an increased ROS-scavenging ability might be beneficial to plant tolerance to abiotic stresses [37]. Conservative cysteine mutations of PDI also affected the activities of antioxidant enzymes and ROS production [47,48]. In this study, the expression levels of ROS scavenging enzyme genes in mutant plants were also much lower than those in WT and OE lines. In addition, the amount of ROS was higher in mutant AtPDI1 plants with conserved cysteines. These results suggested that the catalytic activities of AtPDI1 may be involved in the ROS pathway by regulating the activities of some proteins, but this remains to be further studied.

Conclusion
Based on our findings and previous reports, a hypothetical scheme (Fig. S5) was proposed to depict the molecular basis of AtPDI1 involved in the abiotic stress response. Abiotic stress can cause the accumulation of unfolded and misfolded proteins whose disulphide bonds are disordered and disorganized. AtPDI1 has disulfide isomerase and molecular chaperone activities, which can help these proteins fold correctly. Overexpression of AtPDI1 in Arabidopsis significantly enhanced abiotic stress tolerance. However, mutations of conserved cysteines in the CXXC motif of AtPDI1 decreased the anti-stresses ability of plants, which indicated that the anti-stresses function of AtPDI1 is directly related to the activity of disulfide isomerase.

Construction of plasmids and transformation into Arabidopsis
The open reading frame (ORF) of cysteine-mutant AtPDI1 was amplified by PCR using plasmids previously constructed in our laboratory [28]. Two double mutants (AtPDI1 m1 and AtPDI1m 2 ) of conserved cysteines were amplified using a QuickChange ® II Site-Directed Mutagenesis Kit (Agilent Technologies Inc., US), which changed the cysteines at the motif of Cys128XXCys131 and Cys467XXCys470 of AtPDI1 to Tyr128XXTyr131 and Tyr467XX Tyr470, respectively. Then, all the sequences were inserted into the plasmid of pROKII-GFP containing the CaMV35S promoter by T4 ligase (Fig. S2). The primers used in this investigation are listed in Supplementary Table S1. All recombinant plasmids were confirmed by Shanghai Sunny Biotechnology Co. (Shanghai, China), and the recombinant vectors were transformed into Agrobacterium GV3101 using floral dip transformation. Transgenic plants were selected on the basis of kanamycin resistance and verified by genomic PCR with specific primers (Fig. S3). The homozygous lines were obtained from the T3 generation (Fig. S4).

Plant materials and growth condition
The wild-type Arabidopsis thaliana used in this study was the Columbia ecotype (Col-0) provided by Prof. Chengchao Zheng (Shandong Agricultural University, China), which was originally obtained from Arabidopsis Biological Resource Center (ABRC) [28], and overexpression lines (WT-AtPDI1) were obtained by Zhengrong Zhang as described in our previous study [28]. An AtPDI1 T-DNA insertion knockdown mutant line, pdi (Salk_150463C), was obtained from ABRC and verified by Zhengrong Zhang [28]. Complementary lines (pdi-AtPDI1, pdi-AtPDI1 m1 and pdi-AtPDI1 m2 ) were obtained and identified by Li Yuan in this study as described in the last section, which has not been deposited in a publicly available herbarium. First, the relative mRNA levels of AtPDI1 in the above lines were examined under normal conditions (Fig. S4). Then, the seeds of WT, pdi and transgenic plants were sown on Murashige and Skoog (MS) agar medium in greenhouse conditions at 22 °C under short sunshine conditions with an 8 h light/16 h dark cycle and then transplanted into vermiculite of Shandong Agricultural University under the conditions described by a previous study [28].
Seed germination and root length experiments were performed as described by Zhang et al. (2018) [28]. The washed seeds were germinated on 1/2 MS agar medium with 0.6 mM ABA, 150 mM NaCl, 200 mM mannitol, or 2 mM H 2 O 2 added. Each experiment was carried out in three biological replicates.
For salt or drought stress treatment, 7-day-old wildtype and transgenic plants were irrigated with 250 mM NaCl or 30% (w/v) polyethylene glycol-6000 (PEG-6000) instead of water, respectively. The growth status and the survival rates of 50 seedlings were observed and recorded during the following 21 days. For extreme temperature, 7-day-old wild-type and transgenic plants were treated at − 20 °C for 30 min and then held at room temperature (RT) for 6 h or at 45 °C for 9 h and then subjected to RT for 2 d. Each experiment was carried out in three biological replicates.

RNA extraction and quantitative real-time PCR (qRT-PCR) analysis
Normally grown Arabidopsis plants were treated with solutions containing 150 mM NaCl or with water as a control. The leaves were collected at 6 h, and the total RNA was extracted from the material. Qualified RNA was used to synthesize cDNA with a cDNA synthesis kit (Vazyme, Nanjing, China). The GAPDH gene from Arabidopsis was used as an internal control for expression normalization in qRT-PCR as done in similar investigations, and the relative expression levels were calculated by the 2 − △ △ Ct method [49]. The specific primers used for qRT-PCR are listed in Table S1. The expression levels of ABA-and ROS-related genes in stressed samples (6 h) were compared to the control levels with Student's t test at a significance level of P < 0.05. Each reaction was carried out in three biological replicates.